Biochemistry 2001, 40, 3639-3647
3639
Reaction of Canine Plasminogen with 6-Aminohexanoate: A Thermodynamic Study Combining Fluorescence, Circular Dichroism, and Isothermal Titration Calorimetry† Jack A. Kornblatt,*,‡ Isabelle Rajotte,‡ and Fre´de´ric Heitz§ Enzyme Research Group, Concordia UniVersity, 1455 de MaisonneuVe Ouest, Montreal, Quebec, Canada H3G 1M8, and CRBM, CNRS, 1919, route de Mende, 34293 Montpellier, France ReceiVed August 8, 2000; ReVised Manuscript ReceiVed January 9, 2001
ABSTRACT: The thermodynamics of the binding of 6-aminohexanoate (6-AH) to dog glu-plasminogen has been studied. Fluorescence titrations revealed four binding sites. Three yielded positive fluorescence changes on ligand binding; one yielded a negative fluorescence change. The fluorescence data gave no indication of cooperative interactions. Binding was studied using circular dichroism (CD). Near 295 nm there were small changes associated with binding ligand. These were magnified at 235 nm, a wavelength that is mainly associated with tryptophan bands. The dissociation constants obtained from the fluorescence were applied to the CD data and fit quite well. Below 220 nm, there were no significant differences between samples with or without 6-AH and, therefore, no substantial change in the secondary structure of the protein. Isothermal titration calorimetry was used in combination with the binding constants from fluorescence to study the enthalpy and entropy contributions to 6-AH binding. The enthalpies of association for the four sites are all negative. Their absolute values are small for the tight sites and large for the weakest. -T∆S is negative for the tight sites and positive for the weakest. The binding of 6-AH to plasminogen is entropically driven for the two tightest sites and enthalpically driven for the weakest site. The binding of 6-AH to lys-plasminogen has been studied and differs slightly from binding to glu-plasminogen. Most importantly, the binding of 6-AH for the weak site goes from enthalpy- to entropydriven as is found with the other sites.
Among the leading causes of death in Northern Europe and North America are heart disease and stroke. Fibrin clots that form in the peripheral circulation are dislodged from their sites of origin and move to the heart or brain where they cut off the blood and oxygen supplies; the result is stroke if the clot is in the brain and a myocardial infarct if it is in the heart. Plasminogen, the inactive precursor of the fibrinolytic protein plasmin, gives rise to plasmin and thereby helps to reduce the incidence of these events (1). Plasminogen is a 791-residue peptide composed of seven different and reasonably well characterized domains (2): the N-terminal peptide (NTP)1 of ∼80 amino acids is followed by five kringle domains of 80-90 residues each; these are in turn followed by the carboxy-terminal preproteolytic domain. Intact plasminogen is a closed structure in which lysine 50 of the NTP is noncovalently ligated to a binding † Supported by the Natural Science and Engineering Research Council of Canada. * Corresponding author. Telephone: (514) 848-3404. Fax: (514) 848-2881. E-mail:
[email protected]. ‡ Concordia University. § CNRS. 1 Abbreviations: DPGN, canine plasminogen; glu-plasminogen, fulllength plasminogen consisting of ∼791 amino acids; lys-plasminogen, truncated plasminogen in which residues 1-76, 1-77, or 1-78 have been proteolytically removed; 6-AH, 6-aminohexanoate; NTP, Nterminal peptide; Φ, fluorescence value; CD, circular dichroism. Kringle 1 is the same as site 2, the second tightest 6-AH binding site. Kringle 2 is the same as site 3, the third tightest site. Kringle 3 does not bind 6-AH. Kringle 4 is the same as the tightest binding site, site 1. Kringle 5 is the same as the weakest site (site 4); it is also the binding site for the NTP in the intact protein.
site on kringle 5 (3-5). On the basis of the crystal structures of kringle 5, the site consists of a V-like structure formed from one tryptophan and one tyrosine (6). One extreme of the binding site contains negative charge which can form ion pairs with the positively charged -amino of lysine 50. The crystal structures of microplasmin as well as four of the five kringles have been determined (6-12). Of the five kringles of plasminogen, four (expressed as single domains) will bind the NTP (13) even though only kringle 5 binds it in the intact protein. The four kringles, both in the individually expressed domains and in the intact protein, will also bind 6-AH which presumably resembles lysine 50 or the natural binding sites on the target fibrin. Three of these kringles (1, 2, and 4) bind 6-AH quite tightly; the three have a negative charge in the same position as kringle 5 but also have a positive charge at the other extreme of the binding site. This positive charge forms an ion pair with carboxylate of omega amino acids of the correct length. When plasminogen binds fibrin or 6-AH, lysine 50 is displaced from its position on kringle 5 (3, 4). This produces a large conformational change as the closed form opens; this conformational change is the first step in the activation mechanism to plasmin. The conformational change has been quantified using sedimentation analysis (14-16), circular dichroism (15), fluorescence polarization (17), small-angle X-ray and neutron scattering (18, 19), electron microscopy (20, 21), and gel electrophoresis (22). Plasminogen responds to high hydrostatic pressure which also causes the molecule to open (22). Lysine 50 is displaced
10.1021/bi001857b CCC: $20.00 © 2001 American Chemical Society Published on Web 02/27/2001
3640 Biochemistry, Vol. 40, No. 12, 2001 from its site on kringle 5; the driving force for this is, at least in part, ligand replacement with water and electrostriction of the free amino of lysine 50 instead of ligand replacement with 6-AH. In this report, we show using intrinsic fluorescence, circular dichroism, and isothermal titration calorimetry that the binding of 6-AH by plasminogen can be analyzed as the sum of four hyperbolic functions, each of which represents one of the four binding kringles. The binding constants so obtained have been used to extract the binding entropies and enthalpies from isothermal titration calorimetric experiments. In the intact plasminogen, binding of 6-AH to the two tightest binding kringles is entropically driven. In contrast, binding to the fourth site and the conformational change that results from the binding are enthalpy-driven. This situation applies only in glu-plasminogen, where the NTP is intact. In lysplasminogen, the NTP has been cleaved between positions 77 and 78 (23); there is no longer any ligation between lysine 50 and kringle 5, and any conformational change that takes place when 6-AH binds to kringle 5 is minor compared to the intact protein (4). In lys-plasminogen, the binding of 6-AH to kringle 5 resembles the binding of 6-AH to the other three kringles in that it is largely entropically driven. We note that the vast majority of work currently done on plasminogen is performed on human plasminogen and most of that is done on the recombinant protein expressed in tissue culture cells (4, 24, 25). We have elected to study dog plasminogen because the protein is readily available and easily prepared, because it is free from known human pathogens, because its sequence is homologous to that of the human protein (26, 27), and because it may yield useful crystals for structure determination whereas the human protein has not yet done so. MATERIALS AND METHODS Plasminogen was purified from frozen dog plasma; the approach (28) was to thaw 100 mL of frozen plasma, pass it through a column containing about 1 mL of lysine sepharose, wash with 0.3 M potassium phosphate (Fluka) (pH 7.4) containing 10 µM aprotinin (Sigma), and elute with about 2 mL of 5 mM 6-AH (Fluka) in 0.3 M potassium phosphate (pH 7.4). The sample was then simultaneously concentrated and “dialyzed” with Centricon 30 filters such that the 6-AH was diluted to a maximum of 0.5 µM. Analysis of the dialyzed, concentrated protein showed that it contained less than 0.5 6-AH molecule per DPGN. The final buffer was 25 mM Mes (Boehringer-Mannheim), 25 mM Tris (Sigma), 1 mM magnesium acetate (Fisher), and 0.1 mM EDTA (Fluka) (pH 6.5). The preparation of the purified plasminogen took less than 4 h; the concentration-dialysis step took about the same length of time. The purified plasminogen was divided into aliquots of ∼0.2 mL containing ∼0.2 mg of protein and freeze-dried. The dried powder was kept under vacuum at room temperature and was stable over a period of months. In preparation for fluorescence experiments, the dried powder was reconstituted to the original volume with water and then diluted with the Mes/ Tris/Mg/EDTA buffer. In most instances, it was used within 2 h for the fluorescence titration. On days when we performed more than one experiment, the plasminogen was kept on ice until the second titration was performed. At the end of every titration, a sample was taken for SDS-PAGE
Kornblatt et al. and assayed for cleaved products. If there was any visible formation of lys-plasminogen or plasmin, the data from the titration were discarded. In preparation for CD experiments, the dried powder was reconstituted with 10 mM disodium phosphate and 10 mM monopotassium phosphate (pH 6.8). Lys-plasminogen was prepared from glu-plasminogen. The original purified glu-plasminogen was diluted with 25% glycerol and stored at 253 K for at least six months, during which time it converted to the lys-plasminogen. We could easily distinguish between the glu and lys forms on 7% SDS-PAGE. The glu-plasminogen runs more slowly than phosphorylase b, while the lys-plasminogen runs slightly faster. Our glu-plasminogen preparation contained neither detectable lys-plasminogen, assayed by SDS-PAGE, nor plasmin, assayed by SDS-PAGE and enzymatically with the substrate Chromozym (Boehringer-Mannheim). The lysplasminogen preparation contained >95% lys-plasminogen and no detectable plasmin. Fluorometric titrations were carried out at 288 K using a Shimadzu RF-5000 spectrofluorometer with a built-in stirrer or an Aminco Bowman Series 2 luminescence spectrometer equipped with a built-in stirrer. Samples were contained in standard 4 mL quartz cuvettes that contained a heavy-duty stirring bar designed for cuvettes (VWR catalog no. 58949030, also available from Sigma). Every titration series followed the same procedure. The samples were stirred for 30 s; the stirring was turned off, and the fluorescence was monitored at 280 (excitation) and 360 nm (emission) for 30 s. This was repeated 10 times to assess photodamage. Two microliters of water was added using a Hamilton syringe equipped with a Chaney adapter; the sample was stirred during the addition, the stirring stopped, and fluorescence monitored for 30 s. This was repeated 10 times to further assess photodamage. Over the course of 20 × 30 s, the extent of photodamage was 25 mM. Figure 1B shows the same data but emphasizes the region between 0 and 2 mM 6-AH. The data (b) were fit to the sum of four
Thermodynamics of Plasminogen Binding 6-AH
Biochemistry, Vol. 40, No. 12, 2001 3641
FIGURE 2: Intrinsic fluorescence response of glu-DPGN to the addition of 6-AH. The concentration of DPGN was 2 µM, and the concentration of 6-AH varied between 2 and 500 µM. T ) 288 K; the buffer was the same as that used in Figure 1. The data were fit to the same fluorescence coefficients and dissociation coefficients as used in Figure 1. Black circles are the actual data points; white circles are the fitted points. The inset shows the same data but emphasizes the region between 0 and 20 µM 6-AH.
FIGURE 1: Intrinsic fluorescence response of glu-DPGN to the addition of 6-AH. (A) The concentration of glu-DPGN was 2 µM; the concentration of 6-AH varied from 10 µM to 25 mM. The black circles are the actual data points; the white circles are data fit to the equation shown in the Results. T ) 288 K. The buffer contained 25 mM Mes, 25 mM Tris, 1 mM Mg(OAc)2, and 0.1 mM EDTA (pH 6.5). The fluorescence coefficients and the dissociation coefficients are given in the inset. (B) The same data but emphasizing data between 10 µM and 2 mM 6-AH.
rectangular hyperbolas using the regression function in Sigma Plot and then further refined by simulating the data with the following equation:
Φ ) Φinitial + [[DPGN][6-AH]Φ1/([6-AH] + Kd1)] + [[DPGN][6-AH]Φ2/([6-AH] + Kd2)] + [[DPGN][6-AH]Φ3/([6-AH] + Kd3)] + [[DPGN][6-AH]Φ4/([6-AH] + Kd4)] + ([6-AH]Φ5) where Φ is the fluorescence yield, [DPGN] is the concentration of DPGN, [6-AH] is the concentration of 6-AH, Φ1Φ4 are the fluorescence coefficients associated with the binding of 6-AH to that site, and Kd1-Kd4 are the dissociation coefficients associated with binding to the site. (While the protein contains five kringles, kringle 3 does not bind 6-AH in HPGN and presumably does not do so in DPGN.) Φ5 is a linear term that accounts for very small increases that occur as 6-AH is added; it is independent of DPGN. The global fitting routine allows all sites to be filled at a rate proportional to their respective Kd values. In Figure 1, the black circles are the actual data points while the white circles represent
fitted values. Where, at any given concentration of 6-AH, there is only a white circle, it means that the actual and fitted points are superimposed. The individual fluorescence coefficients and Kd values are shown in the top panel of Figure 1. The fits in panels A and B of Figure 1 are not bad, but they are not perfect. The scatter in the data is such that perfect fits will not be obtained. It is essential to point out that there is no unique fit to the data of Figure 1. With four dissociation constants and four fluorescence coefficients, the data can be approximated with many combinations. We can say that the data were also fit to the sum of two and three hyperbolic functions; the fits were very much poorer. Four-factor Hill plots, three-factor Hill plots, and two-factor Hill plots also gave poor fits over the entire titration range. If it is considered that there are four known binding sites in plasminogen, the use of four sets of coefficients is probably justified (see the Discussion). The experiment whose results are depicted in Figure 1 has been repeated more than 10 times with effectively the same results. Figure 2 shows a fluorescence titration curve using a different sample on a different day. Here we titrated the DPGN with 6-AH but stopped after reaching 0.5 mM. The inset emphasizes the curve between 0 and 20 µM 6-AH. About 70 points were taken in the concentration range from 0 to 0.5 mM. There is scatter in the data; it is especially noticeable in the region around 0.07 mM because the fluorescence changes are so small and because there are many points in this region. The data points (b) were fitted (O) using the fluorescence coefficients and Kd values shown in Figure 1. Once again, the coefficients obtained from the fitting routine give acceptable but not perfect correlations with the actual data. Figure 3 shows the results of similar titrations monitoring the samples with circular dichroism. Panel A demonstrates the changes in the range from 310 to 220 nm, whereas panel
3642 Biochemistry, Vol. 40, No. 12, 2001
Kornblatt et al.
FIGURE 3: (A) Circular dichroic spectra of glu-DPGN from 310 to 220 nm. The solid curve is for glu-DPGN without 6-AH, whereas the dashed curve is for glu-DPGN with 7.6 mM 6-AH. The buffer was 10 mM phosphate (pH 6.8); T ) 288 K. The positive band at 295 nm and the negative band at 235 nm are characteristic of tryptophan. (B) The CD response of glu-DPGN to 6-AH in the far-UV region. The lowest concentration of 6-AH was 0.5 µM, and the concentration was increased in two steps to a final concentration of 7.6 mM. There was no significant difference between any of the curves in the far-UV region. (C) The variation of the CD signal at 235 nm as a function of 6-AH concentration. The data were fit to the dissociation coefficients shown in Figure 1 (s). The ∆ values were evaluated using the same fitting routine as in the fluorescence experiments. Black circles are the data points. The data are shown as a log-linear plot to allow the viewer to see the stepwise nature of the titration.
B concentrates on the far-UV region. There are several significant aspects of the spectra. First is the fact that the tryptophan shoulder at 295 nm, which shows up as a small band with positive ellipticity in the sample without 6-AH, is not present in the sample that contains a high concentration of 6-AH. While the data are not shown, the response is graded over the titration range of from 0 to 25 mM. The major change in the spectrum occurs in the far-UV contribution of tryptophan (220-240 nm). The magnitude of the negative band decreases by a factor of ∼30%; this band is relatively broad and spills over at 225 nm such that there is
a small decrease in intensity at the latter wavelength. In the far-UV region, between 220 and 190 nm, there is absolutely no detectable change in ellipticity resulting from the addition of 6-AH (panel B). Since the absorption bands in the farUV region are primarily the result of secondary structure and ordering of the peptide bond, the conclusion is obvious and must be stated: The very large change in conformation exhibited by DPGN on addition of 6-AH is not the result of large-scale changes in secondary structure. In Figure 3C, we show the results of the titrations of Figure 3A in which the change in ellipticity at 235 nm as a function
Thermodynamics of Plasminogen Binding 6-AH
Biochemistry, Vol. 40, No. 12, 2001 3643 Table 1: Binding of 6-AH to Individual Sites of Canine Plasminogen association constants and thermodynamic constants for binding sites in DPGNa
glu-plasminogen
lys-plasminogen
Ka1 (M-1) ∆G1 (kcal/mol) ∆H1 (kcal/mol) ∆S1 (cal mol-1 K-1) Ka2 (M-1) ∆G2 (kcal/mol) ∆H2 (kcal/mol) ∆S2 (cal mol-1 K-1) Ka3 (M-1) ∆G3 (kcal/mol) ∆H3 (kcal/mol) ∆S3 (cal mol-1 K-1) Ka4 (M-1) ∆G4 (kcal/mol) ∆H4 (kcal/mol) ∆S4 (cal mol-1 K-1)
1.4 × 105 -6.8 -0.45 22 6.6 × 104 -6.4 -0.45 20 1.17 × 103 -4 -2.2 6.3 2.5 × 102 -3.2 -11 -25
2.7 × 105 -7.1 -0.075 25 5 × 104 -6.2 -0.075 21 1.4 × 103 -4.1 -1.5 9.3 2.3 × 102 -3.1 -0.21 10
a The data have been rounded off to two significant figures. As a result, the sum of ∆H and T∆S do not always equal ∆G.
FIGURE 4: Isothermal titration calorimetry of glu-DPGN. The initial concentration of glu-DPGN was 2 µM; the concentration of 6-AH in the injection syringe was ∼500 mM. The temperature was maintained at 288 K. (Top) The actual heat changes occurring during the injection sequence; an expanded view of the data is shown in the inset of the bottom panel. (Bottom) The enthalpic changes as a function of the molar ratio of ligand to protein. The non-zero baseline has been subtracted from the data in the bottom panel. The black squares are the actual heat changes; the solid line is the data fit to the association constants obtained in Figure 1. The calculated enthalpies and entropies are given in Table 1.
of 6-AH concentration has been plotted. Once again, the solid points are the actual data. The data were then fit using the regression function of Sigma Plot to extract values for the extinction coefficients associated with each dissociation coefficient. The solid line represents data fitted on the basis of the dissociation coefficients of Figure 1 in conjunction with the extinction coefficients. As was the case for the fluorescence data, the fits are quite acceptable. The data of Figure 3C are presented as a linear-log plot to emphasize the stepwise nature of the curve (29, 30). Figure 4 shows the results of an isothermal titration calorimetric run in which glu-DPGN was titrated with 6-AH. The top portion of the figure shows the raw heat data from the run. The bottom portion shows the enthalpy changes as a function of the ratio of ligand to plasminogen after subtraction of the sloping baseline. Using the software supplied by Microcal, there was no unique fit to the data. When we applied the dissociation constants of Figure 1 (expressed as association constants) to the data, we obtained the fit shown by the solid line through the data. The enthalpies and calculated entropies are shown in Table 1. The most striking thing about the thermodynamic values is that the enthalpies are all negative but range from -0.5 kcal/mol for the tight sites to -11 kcal/mol for the weakest site. The entropies for
all but the weakest site are positive. -T∆S is evaluated to be approximately -6 kcal/mol for the two tightest sites, -2 kcal/mol for site 3, and 8 kcal/mol for the weakest site. The driving force for 6-AH binding to the weakest site is clearly enthalpy. This result is quite unexpected. Lys-plasminogen lacks the N-terminal peptide containing lysine 50. It is known to exist in an open conformation not identical with that of 6-AH-opened plasminogen but similar (4) (see the Discussion). Fluorescence titrations were carried out using the same protocols as described in the legends of Figures 1 and 2. The data are shown in panels A and B of Figure 5. As was the case with glu-plasminogen, the decrease in fluorescence is followed by an increase. When the data were fit to the sum of four hyperbolas, the fits were virtually as good as those for the glu-plasminogen. The eight constants were not the quite the same in the two preparations; the values for the lys-plasminogen are shown in the figure. The major change occurs in the Kd for the weakest site. It decreases from 4 to 3 mM. The binding of 6-AH for this site is marginally tighter in the lys-plasminogen than in gluplasminogen. The calorimetric response of lys-plasminogen on titration with 6-AH is different from that of the holoenzyme (Figure 6 and Table 1). The experiments whose results are depicted in Figure 6 were similar to those of Figure 4, and the two panels have the same significance. In the lower panel, the line drawn through the data was based on the dissociation constants of Figure 5. As can be seen from the comparison of the two forms of DPGN in Table 1, the first three binding sites do not change substantially. There is a major change that occurs in site 4 associated with kringle 5. ∆S reverses sign and goes from -25 to 10 cal mol-1 K-1. ∆H increases from -11 to -3.3 kcal/mol. This means that the binding of 6-AH to kringle 5, when it is not an exchange reaction, is similar to the binding to the other kringles; the reaction is entropy-driven to the same extent as it is in the other kringles. DISCUSSION In this work, we used fluorescence to detect the binding of 6-AH to canine plasminogen. We previously used this
3644 Biochemistry, Vol. 40, No. 12, 2001
FIGURE 5: (A) Intrinsic fluorescence response of lys-DPGN to 6-AH. The data were evaluated using the equation shown in the Results. The concentration of lys-DPGN was 2 µM; the concentration of 6-AH varied from 10 µM to 25 mM. The black circles are the actual data points; the white circles are data fit to the equation shown in the Results. T ) 288 K. The buffer contained 25 mM Mes, 25 mM Tris, 1 mM Mg(OAc)2, and 0.1 mM EDTA (pH 6.5). The fluorescence coefficients and the dissociation coefficients are given in the inset. (B) The same data but emphasizing data between 10 µM and 2 mM 6-AH.
approach to dissect out the binding constants of human plasminogen (31). The two proteins, canine and human, are very similar. They have tryptophans at identical positions in kringles 4 and 5 and the catalytic domain. The sequences of the other three canine kringles are not known; there are no tryptophans in the NTP (26, 27). We have presented data that relate to the thermodynamics of the interaction of canine plasminogen and 6-aminohexanoate. The final thermodynamic parameters are based entirely on binding constants and a model derived from the fluorescence measurements. The fluorescence data are clear. Fluorescence decreases at 6-AH concentrations between 0 and 0.1 mM 6-AH. At higher concentrations, the fluorescence increases. The fluorescence increase occurs with a slope that indicates that it has a midpoint in the millimolar range. It as been well established by Castellino, Christensen, Markus, and Ponting as well as many other groups that opening of plasminogen occurs when the concentration of 6-AH is above the millimolar range.
Kornblatt et al.
FIGURE 6: Isothermal titration calorimetry of lys-DPGN. The initial concentration of lys-DPGN was 2 µM; the concentration of 6-AH in the injection syringe was ∼500 mM. The temperature was maintained at 288 K. (Top) The actual heat changes occurring during the injection sequence; an expanded view of the data is shown in the inset of the bottom panel. (Bottom) The enthalpic changes as a function of the molar ratio of ligand to protein. The non-zero baseline has been subtracted from the data in the bottom panel. The black squares are the actual heat changes; the solid line is the data fit to the association constants obtained in Figure 5. The calculated enthalpies and entropies are given in Table 1.
The isolated human kringles have been studied. The studies show quite clearly that the kringles containing three tryptophans (kringles 2 and 4) exhibit positive fluorescence changes when they bind 6-AH. Kringles 1 and 5 (each containing two tryptophans) exhibit negative changes when they bind the ligand. A partial explanation for this is given in ref 31. There are at least four kringles in plasminogen that bind 6-AH. Presumably, the fifth kringle also binds but with an affinity that is too low to detect using current methods. The binding of 6-AH to both human (31) and canine plasminogen can be modeled as the sum of four independent binding reactions. We assigned the binding order on the basis of the sign of the fluorescence deflection, the magnitude of the calculated Kd, and the fact that kringle 5 is thought to be the predominant internal ligand for the NTP. However, there are four kringle binding domains, each with its own fluorescence response. There is no unique solution to the binding using only the data obtained from the fluorescence titration. This is the major weakness in our approach. The major strength of the approach is that all the fluorescence, CD, and calorimetric data can be well modeled using the same dissociation coefficients. Our binding order is as follows: kringle 4 binding which is tighter than kringle
Thermodynamics of Plasminogen Binding 6-AH
Biochemistry, Vol. 40, No. 12, 2001 3645
Table 2: Binding of 6-AH to Human and Canine Plasminogena kringle 1 kringle 2 Kd for HPGN 12 µM Kd for DPGN 15 µM Φ for HPGN -6.4 Φ for DPGN -2.1
kringle 3
0.52 mM does not bind 0.85 mM 1 1
kringle 4 kringle 5 4 µM 7 µM 1 1.3
1.6 mM 4 mM 9.2 1.2
a A comparison of the dissociation constants and normalized fluorescence coefficients. The assignment of the dissociation constants and the fluorescence coefficients is based on the sign of the fluorescence change of the isolated human kringles and the approximate dissociation constants of the isolated kringles. The probable reason for the positive and negative changes in fluorescence that occur on ligand binding is described in detail in ref 31. The dissociation constants for the two proteins are quite similar. The relative fluorescence coefficients differ considerably in absolute magnitude but not in sign. Titration of human plasminogen leads to a very large decrease in fluorescence which then recovers to even greater positive values.
1 binding which is tighter than kringle 2 binding which is tighter than kringle 5 binding. Isolated kringle 4 has a positive fluorescence change when binding 6-AH and a very tight binding constant. Kringle 1 has a negative change and a very tight binding constant. Because of the sign of the fluorescence change, we cannot successfully model the binding if we reverse the affinities of kringles 1 and 4. Our assignments of Kd values for kringles 4 and 1 differ from that based on the literature. As isolated, kringle 1 (Kd ) 12 µM) binds 6-AH about twice as tightly as kringle 4 (Kd ) 26 µM). This represents a difference in binding energies of
